In advance by InceptionV3 training model fine - most cerebral sci-film benign and malignant breast image classification are used, the training set accuracy stopped at around 0.88 no longer rise, validation set hovering in 0.6, and the accuracy and validation set loss has a rising trend, there is a fitting? But I used 30 samples to train the network, the network can't convergence, it is how to return a responsibility?

CodePudding user response:

What do you use the data set

CodePudding user response:

Your data set,